Suppose you want to know the effect of the level of capital on firm productivity. Say...

Transcribed Image Text:

Suppose you want to know the effect of the level of capital on firm productivity. Say you have observational firm-level data. This means the data shows you the level of capital and the measure of productivity of each firm. Let's intuitively think of productivity. The firm with higher productivity produces more with the same level of inputs, which are usually capital and labor. Q3.1 Can you think of an omitted variable bias in this context if you only regress capital on firm productivity? You need to clearly explain the two conditions for omitted variable bias. Q3.2 Say there is classical measurement error in the capital variable. Will there be bias? Can you know the direction of the bias? Q3.5 Try to find simultaneous equation bias. Explain what the two equations might be. Q3.3 Say there is classical measurement error in the productivity variable. Will there be bias? Can you know the direction of the bias? Q3.4 Now, say that the data is from a single bank that collects data on firms who applied for loans. If your analysis is about the total US firms, what problem there might be? Suppose you want to know the effect of the level of capital on firm productivity. Say you have observational firm-level data. This means the data shows you the level of capital and the measure of productivity of each firm. Let's intuitively think of productivity. The firm with higher productivity produces more with the same level of inputs, which are usually capital and labor. Q3.1 Can you think of an omitted variable bias in this context if you only regress capital on firm productivity? You need to clearly explain the two conditions for omitted variable bias. Q3.2 Say there is classical measurement error in the capital variable. Will there be bias? Can you know the direction of the bias? Q3.5 Try to find simultaneous equation bias. Explain what the two equations might be. Q3.3 Say there is classical measurement error in the productivity variable. Will there be bias? Can you know the direction of the bias? Q3.4 Now, say that the data is from a single bank that collects data on firms who applied for loans. If your analysis is about the total US firms, what problem there might be?